How to Build a Sound Effects Cue Sheet From Your Manuscript

A sound effects cue sheet is a plain list of every sound your book needs, written against the manuscript before you open the editor. One row per moment: where it happens, what it is, and which kind of audio it is. You can build one for a chapter in under an hour, and the reason to bother is not tidiness. The sheet is what lets you regenerate a chapter six months later, after a line edit, and get the same production back instead of re-deciding every sound from memory.

What a cue sheet is, and what building one does to the prose

Keep the format boring. Chapter, the paragraph or line the sound attaches to, a short description in ordinary words ("rain on a tin roof, low"), and the category. A spreadsheet works. So does a column in the margin of your own draft. What matters is that every entry points at a specific place in the text, because a cue with no anchor becomes a cue you place by feel later.

The read-through that produces the sheet doubles as a close read of your own writing. You find out fast which scenes you actually located and which ones float. You notice when the same storm gets described in four separate paragraphs, each with slightly different weather. Most writers end up editing prose they were not planning to touch.

This guide is about the document. For the wider production picture on a multi-voice piece, our guide on adding sound design to an audio drama covers that side.

Pass one: the sounds the text already names

Read the chapter and write down only what is stated outright. A door slams. A phone rings. Somebody sets a glass down hard enough that it is worth hearing. You are collecting here, so put down everything the prose names even when you suspect you will cut it later.

Do this pass fast. Treating it as a decision-making exercise is what makes people abandon cue sheets halfway through chapter two. The judgment comes later, on a full list, where you can see the density of a scene instead of arguing with yourself one line at a time.

Pass two: implied sounds, and the silences worth protecting

The second pass is the one machines are worst at, and it carries most of the atmosphere. A character steps out of a car and says something bitter. Nothing in that sentence names a sound, but the previous page put them in a downpour, so the line lands inside the rain. Those carried-over conditions are cues, usually written into the sheet as a soundscape that starts earlier and keeps going.

The same pass is where you mark the silences. Some beats get worse with audio under them: a confession, a held breath before an answer, the paragraph after somebody dies. Write those in explicitly as "no cue" rather than leaving the row blank, because a blank row reads as unfinished work and somebody, possibly you, will fill it in. AudioProducer has no control that locks a paragraph against sound, so the sheet is the record of that decision and your job in the editor is to remove whatever lands there.

Auto-Assign Sounds is the machine version of pass one

AudioProducer includes Auto-Assign Sounds, which reads the scene and places music, soundscapes, and one-shot effects from the library automatically. Storms get thunder and wind, transitions get atmosphere. It saves you from hand-picking audio for every paragraph of a long book, and for the sounds your prose states plainly it is quick and mostly right.

Where it is weaker is the second pass. Conditions established several paragraphs earlier are easy for it to drop, since it works from what the scene says rather than from what you were tracking as a writer. It has no way to know which of your quiet moments are deliberate. And it cannot tell that the phone ringing in chapter two is supposed to ring the same way in chapter nine, because that connection lives in your head and nowhere in the text.

So run it, then diff it against your sheet. Auto-Assign is a starting point you review and adjust in the editor, and the sheet turns that review from a vague listen-through into a checklist. Keep what matches, delete what landed on a protected silence, add the carried-over ambience it missed.

Give every cue a category before you open the editor

AudioProducer sorts audio into music beds, ambient soundscapes, and one-shot sound effects, and every row on your sheet should say which one it is. The question that settles it: does this sound have a duration in the story, or does it happen at an instant? A rainstorm outside the window lasts as long as the scene, so it is a soundscape. A struck match happens once, so it is a one-shot. Music is the category you should be most reluctant about, because it comments on a scene rather than existing inside it.

Categorizing on paper saves time in the Sounds panel, and it makes density visible: three soundscapes stacked in one scene is a mix problem you can catch before rendering anything. For the specifics of each type, see our guides on adding ambient sound to an audio story and adding background music to an audiobook.

Prune the sheet, then keep it

Now go back over the finished list and cut. The test that works: if removing a cue changes nothing about where the listener thinks they are or what just happened, it is decoration. Our guide on sound effects and music in an audiobook goes further into where effects clutter, and a sheet is much easier to argue with than a finished render.

When the audio is generated, do not throw the sheet away. Files are saved in your account and each chapter can be downloaded separately, but the sheet is the part that makes the production repeatable. Edit a paragraph, regenerate, and you rebuild against a document instead of a memory. If the book is part of a series, the recurring locations are already written down, so the next book starts with a sheet rather than a blank page. Pair it with a listening pass before release, as in our guide to quality-checking an AI audiobook, and with the per-line emotion settings if your dialogue needs shaping too.

You can try the whole flow on the free tier, which includes 1,200 words at no cost and no card, with paid plans from $39.99 per month. Bring the manuscript in as an EPUB or paste it directly, since those are the two import paths we support. The output is an MP3 you download and publish wherever you already publish; we do not distribute or list audio for you. Voice cloning is available for a voice you own or have permission to use.